ModelsObject Detection
YOLO-NAS
Object Detection
Deploy with Roboflow

YOLO-NAS

YOLO-NAS is an object detection model developed by Deci that achieves SOTA performances compared to YOLOv5, v7, and v8.
Deci AI
•
GitHubGitHub
•
star
4.4k+
stars
•
Apache 2.0
license
•
Developed by
Deci AI
•
Released
May 3, 2023
•
Arxiv logo markPaper

Try the Model

Use the widget below to experiment with YOLO-NAS. You can detect COCO classes such as people, vehicles, animals, household items.

Overview

YOLO-NAS is an object detection model developed by Deci that achieves SOTA performances compared to YOLOv5, v7, and v8. YOLO-NAS is pre-trained on multiple prominent datasets including COCO, Objects365, and Roboflow 100. This overachieving pre-training ensures its precision amongst numerous tasks. Try training YOLO-NAS on your own dataset.

YOLO-NAS performance (Source: Deci YOLO-NAS README).

Use Cases

video surveillance

medical diagnosis

wildlife monitoring

real-time object detection tasks

‍

YOLO-NAS License

YOLO-NAS

is licensed under a 

Apache 2.0

license. 

Performance

Model mAP Latency (ms)
YOLO-NAS S 47.5 3.21
YOLO-NAS M 51.55 5.85
YOLO-NAS L 52.22 7.87
YOLO-NAS S INT-8 47.03 2.36
YOLO-NAS M INT-8 51.0 3.78
YOLO-NAS L INT-8 52.1 4.78

mAP numbers in table reported for Coco 2017 Val dataset and latency benchmarked for 640x640 images on Nvidia T4 GPU.

Deploy a YOLO-NAS API

You can use Roboflow Inference to deploy a

YOLO-NAS

API on your hardware. You can deploy the model on CPU (i.e. Raspberry Pi, AI PCs) and GPU devices (i.e. NVIDIA Jetson, NVIDIA T4).

Below are instructions on how to deploy your own model API.

First, install Inference:‍


pip install inference

To try a demo with a model trained on the Microsoft COCO dataset, use:


import inference
model = inference.load_roboflow_model("yolo-nas-s-640")
results = model.infer(image="YOUR_IMAGE.jpg")

Above, replace:

  • YOUR_IMAGE.jpg with the path to your image.

You can also run fine-tuned models with Inference.

Retrieve your Roboflow API key and save it in an environment variable called ROBOFLOW_API_KEY:


export ROBOFLOW_API_KEY="your-api-key"

To use your model, run the following code:


import inference
model = inference.load_roboflow_model("model-name/version")
results = model.infer(image="YOUR_IMAGE.jpg")

Above, replace:

Label Data Automatically with YOLO-NAS

You can automatically label a dataset using YOLO-NAS with help from Autodistill, an open source package for training computer vision models. You can label a folder of images automatically with only a few lines of code. Below, see our tutorials that demonstrate how to use YOLO-NAS to train a computer vision model.

YOLO-NAS Alternatives

Compare YOLO-NAS Model Sizes

No items found.
Text Link
Deploy YOLO-NAS in One Click
Explore what YOLO-NAS can do in a fully interactive workflow. GPU-backed, no code required.
Deploy Now with free GPU
Rocket

Deploy to Production

Roboflow offers a range of SDKs with which you can deploy your model to production.
Deploy to...
Rocket

Build a Workflow

Use YOLO-NAS to Build Computer Vision Pipelines and Applications
Images

Data

Augmentation GuideLabeling Guide
Flask

Training

Model OverviewTrain on ColabTrain on JupyterTrain on KaggleTrain on SageMakerSee Model AlternativesView Benchmarks
Screwdriver and Wrench

Utilities

Plot PredictionsFilter PredictionsBlur PredictionsCreate A Confusion MatrixUse SAHI for Small Object Detection
Code Compare

Compare to Related Models

Curious about how this model compares to others? Check out our model comparisons.
Compare with...
arrow progress

Convert Annotation Format

YOLOv8 uses the uses the YOLOv8 PyTorch TXT annotation format. If your annotation is in a different format, you can use Roboflow's annotation conversion tools to get your data into the right format.
Convert Data Between Formats

Architecture

YOLO

Frameworks

PyTorch

Annotation Format

YOLOv5 PyTorch TXT

Popular YOLO-NAS Use Cases

No items found.

YOLO-NAS Integrations

YOLO-NAS to Active Learning Data Collector

Build a computer vision workflow that connects YOLO-NAS to Active Learning Data Collector.

YOLO-NAS to Kafka Publish

Build a computer vision workflow that connects YOLO-NAS to Kafka Publish.

YOLO-NAS to CSV Sink

Build a computer vision workflow that connects YOLO-NAS to CSV Sink.

YOLO-NAS to Azure Sink

Build a computer vision workflow that connects YOLO-NAS to Azure Sink.

YOLO-NAS to Webhook Publish

Build a computer vision workflow that connects YOLO-NAS to Webhook Publish.

YOLO-NAS to Super Annotator

Build a computer vision workflow that connects YOLO-NAS to Super Annotator.

YOLO-NAS to SMS

Build a computer vision workflow that connects YOLO-NAS to SMS.

YOLO-NAS to S3 Upload

Build a computer vision workflow that connects YOLO-NAS to S3 Upload.

YOLO-NAS to MQTT Publisher

Build a computer vision workflow that connects YOLO-NAS to MQTT Publisher.

Roboflow Keypoint Detection Model to YOLO-NAS

Build a computer vision workflow that connects Roboflow Keypoint Detection Model to YOLO-NAS.

YOLO-NAS to YOLOv5 Instance Segmentation

Build a computer vision workflow that connects YOLO-NAS to YOLOv5 Instance Segmentation.

YOLO-NAS to YOLO-NAS

Build a computer vision workflow that connects YOLO-NAS to YOLO-NAS.

YOLO-NAS to YOLOv9

Build a computer vision workflow that connects YOLO-NAS to YOLOv9.

YOLOv8 to YOLO-NAS

Build a computer vision workflow that connects YOLOv8 to YOLO-NAS.

YOLOv7 Instance Segmentation to YOLO-NAS

Build a computer vision workflow that connects YOLOv7 Instance Segmentation to YOLO-NAS.

YOLOv9 to YOLO-NAS

Build a computer vision workflow that connects YOLOv9 to YOLO-NAS.

YOLO-NAS to YOLOv8 Classification

Build a computer vision workflow that connects YOLO-NAS to YOLOv8 Classification.

YOLO-NAS to YOLOv8

Build a computer vision workflow that connects YOLO-NAS to YOLOv8.

YOLOv8 Keypoint Detection to YOLO-NAS

Build a computer vision workflow that connects YOLOv8 Keypoint Detection to YOLO-NAS.

YOLOv5 Instance Segmentation to YOLO-NAS

Build a computer vision workflow that connects YOLOv5 Instance Segmentation to YOLO-NAS.

YOLOv8 Classification to YOLO-NAS

Build a computer vision workflow that connects YOLOv8 Classification to YOLO-NAS.

YOLO-NAS to Roboflow Keypoint Detection Model

Build a computer vision workflow that connects YOLO-NAS to Roboflow Keypoint Detection Model.

YOLOv8 Instance Segmentation to YOLO-NAS

Build a computer vision workflow that connects YOLOv8 Instance Segmentation to YOLO-NAS.

Roboflow Object Detection Model to YOLO-NAS

Build a computer vision workflow that connects Roboflow Object Detection Model to YOLO-NAS.

YOLO-NAS to YOLOv8 Keypoint Detection

Build a computer vision workflow that connects YOLO-NAS to YOLOv8 Keypoint Detection.

YOLO-NAS to Roboflow Object Detection Model

Build a computer vision workflow that connects YOLO-NAS to Roboflow Object Detection Model.

Roboflow Instance Segmentation Model to YOLO-NAS

Build a computer vision workflow that connects Roboflow Instance Segmentation Model to YOLO-NAS.

YOLO-NAS to Roboflow Classification Model

Build a computer vision workflow that connects YOLO-NAS to Roboflow Classification Model.

YOLO-NAS to Roboflow Instance Segmentation Model

Build a computer vision workflow that connects YOLO-NAS to Roboflow Instance Segmentation Model.

Explore This Model on Roboflow

Train YOLO-NAS - SOTA Object Detection Model - on Custom Dataset
Video
Video

Train YOLO-NAS - SOTA Object Detection Model - on Custom Dataset

Learn how to use YOLO-NAS in our comprehensive video guide.
How to Train YOLO-NAS on a Custom Dataset
Blog
Blog

How to Train YOLO-NAS on a Custom Dataset

YOLO-NAS is the latest state-of-the-art real-time object detection model. Learn how to train YOLO-NAS on your custom data.
Ready to get started?

Join over 1 million developers building with Roboflow

Try it for FreeTalk to Sales